The Data Science Career Guides series brings you tips for starting, advancing, and enjoying your data science career. Today, we’re sharing Coursera Data Science Manager Dan Saber’s advice for landing your very first data science job. Dan’s piece originally appeared on Inside Big Data.
It’s a familiar dilemma. You’ve done your research, read some books, taken some online classes – and at long last, you’re finally ready to get real-life work experience as a Data Scientist.
But as you browse job postings, you become discouraged: “They want me to be a d3 Expert and a Deep Learning ‘Ninja’? A ‘Wizard’ of ETL and a tidyverse-loving #Rstats ‘Samurai’? What does it even mean to be a Viking of scikit-learn by ascending to the Valhalla of XGBoost? Is that, like, two years of work experience? Three?”
As you add a few more classes to your ever-expanding study plan, you let out a sigh: “Maybe Data Science can be my second career when I retire.”
While I believe that lifelong learning is the universe’s most powerful force for good, it’s easy to overdo. If you’ve developed a solid foundation, you may be ready to look for your first job – even if your resume doesn’t perfectly match the requirements for your dream position. This is particularly true in Data Science, where no single human being will ever be qualified for your average entry-level job posting.
And while we’re not going to fix the Data Science hiring market in a single article, I can at minimum offer my humble perspective on how to get genuine Data Science experience on your resume.
1. Build a portfolio
This advice almost certainly counts as cliched; however, like many great cliches, it contains some profound truth if you can keep yourself from rolling your eyes.
Let’s put ourselves in the shoes of a hypothetical Data Science hiring manager. We’re juggling a handful of brilliant-but-unruly direct reports and a dozen or so crunch-time projects. Last week, the boss walked in with great news – we’re going to lead a big, important new initiative! And sure, we can hire someone to help. So now we’re staring at your job application, trying to figure out if you’re interested and competent – and we need to do it quickly.
Ah, but you’ve included a portfolio. We open it and see a few well-thought-out projects that demonstrate a range of skills. Anyone who took time to put this together is interested. Anyone who could put this together is probably competent. We’re intrigued by a particular project. We have a few questions. Could you come in tomorrow for an interview?
The above is a dramatization, of course, but I will always – 100 times out of 100 – browse through somebody’s Github or blog if they link to it. Moreover, you’d be surprised by how few people do this; it really does differentiate you, and can actually be almost as impactful a signal as previous Data Science experience. As discussed above, “Data Science” encompasses an enormous range of skills, so many hiring managers are looking more for evidence that you can learn new skills and apply them to real-life problems than they are for a laundry list of certifications.
A portfolio isn’t just a tool for hiring managers, either. Building a portfolio has profound personal benefits, including confidence (being able to speak to real things you’ve done is good); genuine skill development (reading about things is never the same as actually doing things); and what Bill Burnett and Dave Evans call “prototyping your life,” which in this case means figuring out if Data Science is something you really want to do by experimenting with it in a lightweight, low-cost way.
2. Build upon your present circumstances
Though it’s existed in various forms for some time, what we call “Data Science” is still a new field. Nearly everyone who calls themselves Data Scientists got to where they are by making a transition from a more traditional career. For some – e.g., a statistician or a PhD in a technical field – the transition may have been straightforward.
For others, however, it will have been more involved. I, for example, worked in finance: a data-rich field, but certainly not one that had embraced all of what Data Science had to offer. However, I saw this gap as an opportunity to improve both my field and my skills. By night, I took online classes (like Andrew Ng’s Machine Learning class on Coursera), and by day, I applied what I learned to bring better tools or fresh perspective to my more quantitative projects. This enabled me to develop real-life expertise.
So ask yourself: are there aspects of your current job that would benefit from a bit of Data Science magic? And if not, could you volunteer your burgeoning data skills for a non-profit, political campaign, or open source project? If so, start building experience now on a project-by-project basis. Even if your official job title doesn’t change, those projects are great resume builders, help beef up a portfolio, and are great stories to discuss in interviews to demonstrate you’re a true Data Science practitioner
3. Consider a Master’s program or bootcamp
Portfolios and projects are wonderful, but sometimes you need immersive experience to take you the last mile. In my case, I went to a Data Science bootcamp, which allowed me to accelerate my skill development considerably; the bootcamp experience also gave me a cogent narrative for my transition from finance into Data Science. (Never underestimate the power of being able to construct a narrative for how you got from Point A to Point B.)
The market for Data Science bootcamp-level curriculums has only grown stronger and more competitive in the past few years. In addition to a variety of in-person immersive options, there are also robust online programs – ranging from certificates built on series of courses to full online Data Science Master’s programs. While this option isn’t for everyone, I’m glad I pursued it: after all, if you’ve done enough prototyping, the next step is to go out and build something.